We're hiring

AI Delivery Orchestrator

We don't write much code anymore. We orchestrate AI that does. We're looking for the person who can run that.

Agent mode · new

Evaluate this role with your own AI

Point your ChatGPT or Claude at this page. Because it already knows you, it’ll tell you—honestly—whether you’d be a fit, then help you prepare and apply. This is how we work, so it’s how we hire: work it through with AI.

The “Open in…” buttons copy the prompt and open a new chat — just paste (⌘/Ctrl+V).

Orange Hill is a small, senior, profitable software studio that builds and maintains platforms for high-profile clients — the Serbian Chamber of Commerce, m:tel Austria, Eucerin, Coca-Cola, and others around the world — across Serbia and the DACH region. Roughly two-thirds of our delivery is already automated. We're not hiring someone to type faster. We're hiring someone to direct.

What you'll own

  • Run several client projects in parallel — by orchestrating AI agents and steering them to a high-quality result, not by coding each one by hand.
  • Own delivery end to end: architecture decisions, the path to production (DevOps, servers, integrations), and the quality bar.
  • Be the taste in the room — judge what's good, what's well-planned, what's genuinely good UX, and kill what isn't.
  • Turn vague client goals into shippable products, fast. You'll run many things at once — but you know judgment doesn't parallelize, so you protect it.

Who you are

  • Ruthlessly organized. You can hold several projects in your head and keep every one moving. This is the trait we care about most — it's what makes parallelism and agent orchestration possible.
  • A real engineer. You've built and shipped software, you understand servers and DevOps, and you know your way around third-party APIs. You can judge code even when you're not the one writing it.
  • Creative, with taste. You believe the best builders are part scientist, part artist. You have opinions about design and UX — you don't have to be a designer, but you know good from bad and you care.
  • AI-native by default. Your instinct is “what's the best way to point AI at this,” not “let me hand-code it.” You're genuinely excited about where this is going.

This might not be your role if…

  • You love grinding one hard problem in deep solo focus and don't want to let go of the keyboard. Great engineer — wrong seat.
  • You treat AI tools as a threat or a toy.
  • You're brilliant but disorganized. Organization is the non-negotiable here; we can't coach it fast enough.
  • You want a narrow IC lane with someone else owning the messy whole.

How we work

Agents do the building; humans hold the judgment. We describe and verify; we don't type. We run many parallel threads but protect the one serial thing — good decisions — with deliberate thinking time. We expect you to recognize what's good and delegate execution to AI, not to program the traditional way. If that sounds like a downgrade, this isn't your role. If it sounds like leverage, keep reading.

How we think

We write about all of this publicly. If these resonate, you'll fit: Judgment Doesn't Parallelize, Software Development Is Dead, and Affinity, Style, Taste.

What we offer

  • Real autonomy, ownership, and upside — high-leverage work with a senior team behind you (base + performance).
  • A front-row seat building our own AI-first products, not just client work.
  • A team that's small, senior, and allergic to bureaucracy.
  • A path that grows with our DACH / Austria expansion.

How to apply

Don't send just a CV. We'll ask you to do one small thing: read an anonymized snapshot of a real Orange Hill week — three live projects, one of them on fire — and write your one-page plan for it: what you'd do yourself, where you'd point AI agents, what you'd let wait, and how you'd keep nothing from slipping.

The fastest way to do it well is the AI you just connected above — it has the full role and can work the task through with you. That's not a shortcut; it's the job.

Agent mode · new

Evaluate this role with your own AI

Point your ChatGPT or Claude at this page. Because it already knows you, it’ll tell you—honestly—whether you’d be a fit, then help you prepare and apply. This is how we work, so it’s how we hire: work it through with AI.

The “Open in…” buttons copy the prompt and open a new chat — just paste (⌘/Ctrl+V).

Submit your application

Worked it through with your AI? Paste it in. We reply to everyone who does the task.

CV / résumé (required)
PDF, DOC or DOCX · max 10 MB
Links (optional)